• The Journal of the Korea institute of electronic communication sciences
• /
• v.15 no.6
• /
• pp.1209-1216
• /
• 2020

In this study, we propose a method for detecting coastal surface wastes using an UAV(Unmanned Aerial Vehicle) remote sensing method and an object detection algorithm based on deep learning. An object detection algorithm based on deep neural networks was proposed to detect coastal debris in aerial images. A deep neural network model was trained with image datasets of three classes: PET, Styrofoam, and plastics. And the detection accuracy of each class was compared with Darknet-53. Through this, it was possible to monitor the wastes landing on the shore by type through unmanned aerial vehicles. In the future, if the method proposed in this study is applied, a complete enumeration of the whole beach will be possible. It is believed that it can contribute to increase the efficiency of the marine environment monitoring field.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.25 no.4
• /
• pp.151-162
• /
• 2022

Recently, the forests in Korea have accumulated damage due to continuous forest disasters, and the need for technologies to monitor forest managements is being issued. The size of the affected area is large terrain, technologies using drones, artificial intelligence, and big data are being studied. In this study, a standard dataset were conducted to develop an algorithm that automatically detects suspicious trees damaged by forest pests using deep learning and drones. Experiments using the YOLO model among object detection algorithm models, the YOLOv4-P7 model showed the highest recall rate of 69.69% and precision of 69.15%. It was confirmed that YOLOv4-P7 should be used as an automatic detection algorithm model for trees suspected of being damaged by forest pests, considering the detection target is an ortho-image with a large image size.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.27 no.1
• /
• pp.8-14
• /
• 2023

The existing self-supervised learning-based CutPaste generated synthetic data by cutting and attaching specific patches from normal images and then performed anomaly detection. However, this method has a problem in that there is a clear difference in the boundary of the patch. NSA for solving these problems have achieved higher anomaly detection performance by generating natural synthetic data through Poisson Blending. However, NSA has the disadvantage of having many hyperparameters that need to be adjusted for each class. In this paper, synthetic data similar to normal were generated by a simple method of making the size of the synthetic patch very small. At this time, since the patches are so locally synthesized, models that learn local features can easily overfit synthetic data. Therefore, we performed anomaly detection using gMLP, which learns global features, and even with simple synthesis methods, we were able to achieve higher performance than conventional self-supervised learning techniques.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.33 no.3
• /
• pp.411-425
• /
• 2023

Today, AI (Artificial Intelligence) technology is widely used in various fields, performing classification and regression tasks according to the purpose of use, and research is also actively progressing. Especially in the field of security, unexpected threats need to be detected, and unsupervised learning-based anomaly detection techniques that can detect threats without adding known threat information to the model training process are promising methods. However, most of the preceding studies that provide interpretability for AI judgments are designed for supervised learning, so it is difficult to apply them to unsupervised learning models with fundamentally different learning methods. In addition, previously researched vision-centered AI mechanism interpretation studies are not suitable for application to the security field that is not expressed in images. Therefore, In this paper, we use a technique that provides interpretability for detected anomalies by searching for and comparing optimization references, which are the source of intrusion attacks. In this paper, based on reference, we propose additional logic to search for data closest to real data. Based on real data, it aims to provide a more intuitive interpretation of anomalies and to promote effective use of an anomaly detection model in the security field.

• Proceedings of the Korean Information Science Society Conference
• /
• 2010.11a
• /
• pp.127-128
• /
• 2010

본 논문에서는 CCTV 감시 영상에서 취득한 얼굴 이미지를 이용하여, 범죄자 감시목록에 등록된 범죄 용의자를 탐지 식별하는 시스템을 설계 및 구현하였다. 특히 본 논문에서 제안한 SVDD와 SRC를 혼합한 계층적 구조의 범죄 용의자 식별 모듈은 다음과 같은 특성을 갖는다: 1) 먼저 SVDD를 이용하여 범죄 용의자만을 빠르게 인식함으로써, 일반인에 대한 불필요한 범죄자 식별 연산을 수행하지 않는다; 2) 다양한 식별 성능을 저해하는 환경에서도 이미 강인한 성능이 검증된 SRC를 범죄 용의자 식별과정에 적용함으로써 안정적이고 정확한 식별 시스템을 보장한다; 3) 동일 생체 특정의 반복적 사용을 통한 다수결 투표전략을 취함으로써 시스템의 신뢰도를 보장한다; 4) 점증적 갱신의 학습 능력으로 인하여 범죄 용의자 감시목록 데이터베이스의 변화에도 능동적으로 적응한다 실제 KUFD(Korea University Face Database)를 자체 제작하고 캠퍼스 내에서 CCTV 환경의 얼굴 인식 기반 범죄 용의자 탐지 및 식별 시스템 환경을 모의 구축하여 실험적으로 제안된 시스템의 성능을 검증한다.

• Journal of Korea Society of Industrial Information Systems
• /
• v.23 no.4
• /
• pp.41-53
• /
• 2018

Recently, the development of machine learning technology has led to the application of deep learning technology to existing image based application systems. In this context, some researches have been made to apply CNN (Convolutional Neural Network) to the field of fire detection. To verify the effects of existing preprocessing and feature extraction methods on fire detection when combined with CNN, in this paper, the recognition performance and learning time are evaluated by changing the VGG19 CNN structure while gradually increasing the convolution layer. In general, the accuracy is better when the image is not preprocessed. Also it's shown that the preprocessing method and the feature extraction method have many benefits in terms of learning speed.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06b
• /
• pp.55-57
• /
• 2006

본 논문에서는 컬러코드를 이용하여 증강현실 시스템에 사용 가능한 태그를 탐지하는 알고리즘을 설계하고 차단 현상에 강인한 알고리즘을 제안하였다. 기존의 ARToolkit에서 태그의 일부분이 사용자 또는 다른 물체에 의해 가려지게 될 경우 증강되었던 객체가 순간 사라져 버리는 불안정성 (Instability) 문제를 해결하기 위한 방법에 초점을 맞춘다. 불안정성의 문제는 이미지 안에 태그가 존재하지만 해당하는 객체를 증강시키지 못하는 False Negative 에러와 태그가 존재하지 않는 곳에 잘못된 객체를 증강시키는 False Positive 에러로 분류 될 수 있다. 제안된 탐지 알고리즘으로 특정 컬러 영역을 분리하여 모서리 여부를 판별하고 모서리인 경우 가려진 꼭지점의 위치를 추출하여 태그가 차단에 의하여 가려졌을 때에도 객체를 안정적으로 증강시킬 수 있다. 기존 AR 시스템들의 태그를 가지고 Daylight 65, Illuminant A. CWF, TL84의 4가지의 표준 조명하에 컬러코드 4종류, ARToolkit 태그 4개, ARTag 4개를 이용하여 실험을 진행하여 차단 현상이 발생하면 전혀 객체를 증강시킬 수 없었던 ARToolkit에서도 DayLight65의 경우 50%의 False Negative. False Positive rate을 보여 기존 증강현실 시스템에서 보였던 불안정성 문제를 개선하였다.

• Convergence Security Journal
• /
• v.24 no.1
• /
• pp.77-82
• /
• 2024

Recent research has seen significant improvements in various object detection and classification models overall. However, the study of object detection and classification in situations where objects are partially obscured remains an intriguing research topic. Particularly in the military domain, unmanned combat systems are often used to detect and classify objects, which are typically partially concealed or camouflaged in military scenarios. In this study, a method is proposed to enhance the classification performance of partially obscured objects. This method involves adding occlusions to specific parts of object images, considering the surrounding environment, and has been shown to improve the classification performance for concealed and obscured objects. Experimental results demonstrate that the proposed method leads to enhanced object classification compared to conventional methods for concealed and obscured objects.

• Journal of Korea Society of Industrial Information Systems
• /
• v.12 no.3
• /
• pp.104-115
• /
• 2007

As soon as an intrusion is detected by kernel backdoor, the proposed method can be preserve secure and trustworthy evidence even in a damaged system. As an experimental tool, we implement a backup and analysis system, which can be response quickly, to minimize the damages. In this paper, we propose a method, which can restore the deleted log file and analyze the image of a hard disk, to be able to expose the location of a intruder.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2018.05a
• /
• pp.439-441
• /
• 2018

딥 러닝의 높은 성능으로 여러 분야에 사용되며 기침 탐지에서도 수행된다. 이 때 기침과 유사한 재채기, 큰 소리는 단일 데이터만으로는 구분하기에 한계가 있다. 본 논문에서는 기존의 오디오 데이터와 오디오 데이터를 인코딩 한 스펙트로그램 이미지 데이터를 함께 학습하는 멀티 모달 딥 러닝을 적용하는 방법을 사용한다.